EVIEWS操作各种模型学习.docx
《EVIEWS操作各种模型学习.docx》由会员分享,可在线阅读,更多相关《EVIEWS操作各种模型学习.docx(50页珍藏版)》请在冰豆网上搜索。
EVIEWS操作各种模型学习
章、图形基础与回归
原始数据
UR
URIR
UR
URIR
1978
0.15
2.97
1997
0.31
2.22
1979
0.16
2.32
1998
0.32
2.22
1980
0.17
2.43
1999
0.34
2.35
1981
0.17
2.09
2000
0.40
2.43
1982
0.18
1.68
2001
0.41
2.48
1983
0.18
1.70
2002
0.41
2.77
1984
0.18
1.50
2003
0.42
2.85
1985
0.23
1.67
2004
0.43
2.77
1986
0.24
1.91
2005
0.43
2.83
1987
0.25
2.06
2006
0.44
2.86
1988
0.27
2.26
2007
0.44
2.87
1989
0.28
2.20
2008
0.45
2.82
1990
0.29
2.12
2009
0.46
2.85
1991
0.26
2.54
2010
0.50
2.89
1992
0.29
2.76
2011
0.52
2.84
1993
0.31
3.11
2012
0.54
2.91
1994
0.28
2.85
2013
0.55
2.91
1995
0.31
3.48
2014
0.55
2.83
1996
0.34
2.33
,、散点图
RRlu
-eE」ONo
:
、分布图:
JB>3判断为正太分布
Series:
UR
Sample137
Observations37
Mean
0.336145
Median
0.312400
Maximum
0.547000
Minimum
0.15087-
Std.Dev.
0.120626
Skewness
0.143256
Kurtosis
1.887565
Jarque-Bera
2.034384
Probability
0.361609
三、UR的单因素联表
TabulationofUR
Date:
09/05/15Time:
21:
25
Sample:
137
Includedobservations:
37
Numberofcategories:
5
Value
Count
Percent
Cumulative
Count
Cumulative
Percent
[0.1,0.2)
7
18.92
7
18.92
[0.2,0.3)
9
24.32
16
43.24
[0.3,0.4)
6
16.22
22
59.46
[0.4,0.5)
11
29.73
33
89.19
[0.5,0.6)
4
10.81
37
100.00
Total
37
100.00
37
100.00
四、协方差与相关性
CovarianceAnalysis:
Ordinary
Date:
09/05/15Time:
21:
40
Sample:
137
Includedobservations:
37
Covariance
Correlation
UR
URIR
UR
0.014157
1.000000
URIR
0.033170
0.204852
0.615934
1.000000
Date:
09/05/15Time:
21:
44
Sample:
137
Includedobservations:
37
Correlationsareasymptoticallyconsistentapproximations
URrURlR(-i)
UR,URIR(+i>
lag
lead
Oi
0
0.6159
0.&159
1
0.5767
0.6549
2
0.4997
0.6515
3
0.4282
0.6384
4
0.3795
0.5755
5
0.3621
0.4990
6
0.3267
0.4159
7
0.3292
0.2918
8
0.2983
0.1732
9
0.2585
0.0795
10
0.1997
0.0124
11
0.1420
-0.0556
12
□0809
-0.0998
13
0.0097
-0.1538
14
-0.0231
■0.1740
15
-0.033B
-0.1790
16
-0.0565
*0.1401
17
-0.0433
-0.1238
18
-0.0316
-0.0532
19
-0.0395
-0.0939
20
-0.1629
-0.1381
五、CDF经验分布图
六、Q-Q图
URIR
QuantilesofURIR
七、回归散点图
UR
邻近拟合散点图:
(分布回归的结果)
八、实际值、拟合值、残差值折线图
ResidualActualFitted
9
StandardizedResiduals
九、回归模型预测
URF—±2S.E
Forecast:
URF
Actual:
UR
Forecastsample:
137
Includedobservations:
37
RootMeanSquaredError0.093736
MeanAbsoluteError0.072711
MeanAbs.PercentError25.93140
TheilInequalityCoefficient0.133790
BiasProportion0.000000
VarianeeProportion0.237674
CovarianeeProportion0.762326
十、两回归系数的联合检验置信区间是一个椭圆区域
C
(1)
!
^一、Wald系数约束条件检验
WaldTest:
Equation:
Untitled
TestStatistic
Value
df
Probability
F-statisticChi-square
272.1503
272.1503
(1.35)
1
0.0000
0.0000
NullHypothesisSummary:
NormalizedRestriction[=Q)
Value
Std.Err,
-1+C
(1)+C
(2)
-0,907527
0.055012
RestrictionsarelinearincgeAI匚ients
Chow分割点检验结果
ChowBreakpointTest1991
NullHypothesis:
Nobreaksatspecifiedbreakpoints
Varyingregressors:
Allequationvariables
EquationSample:
19782014
F'Statistic
11.00551
Prob.F(2r33)
0.0002
Loglikelihoodratio
18.90797
Prob,Chi-Square
(2)
0.0001
WaldStatistic
22.01103
Prob.Chi-Squa(e
(2)
0.0000
F、LR的P值显著,表示:
模型无显著的结构变化
十二、Chow稳定性检验(p75)
Chow预测结果:
ChowForecastTest:
Forecastfrom1991to2014
F-statistic
Loglikelihoodratio
4.744782
89.83840
PfOb.F(24,11)
Prob.Chi-Square(24)
0.0050
0.0000
TestEquation:
DependentVariable:
URMethod:
LeastSquares
Date:
09/07/15Time:
10:
31
Sample:
197S1990
Included!
observations:
13
Coefficient
Std.Errort-Statistic
Prob.
URIR
C
-0014716
0.241289
0.037436-0393094
0.0787753.063016
07018
0.0100
R-squared
AdjustedR-squared
S.Eofregression
SumsquaredresidLoglikelihoodF-statistic
Prob(F-statistic)
0.013853
*0.075797
0.051024
0.028637
21.32073
0.154523
0.701762
Lieandependentvar
S.Ddependentvar.Akaikeinfocriterion
SchwarzcriterionHannan-Ouinncriter.
Durbin-Watsonstat
0.210827
0.049193
2972421-2.885505-2990286
0.127458
十三、零均值附近的递归残差曲线图
2
2•递归OLS的CUSUM检验曲线图(p78)
CUSUM5%Significance
*注:
红线为5%的临界值线,在1991年后的CUSUM曲线变得十分陡峭,说明:
回归方程系数并不是稳定的。
3.—步预测检验:
N-StepProbabilityRecursiveResiduals
One-Step
Probability
Recursiv
eResiduals
*注:
上部分是递归残差,下部分是检验显著性的概率
.3
RecursiveC⑴Estimates士2S.E.
RecursiveC
(2)Estimates士2S.E.
十四、White异方差检验
Obs*R-squared=10.4,其P值=0.0055表示残差存在异方差性。
F统计量表示:
检验辅助方程的整体显著性,下图中整体显著。
HeteroskedasticityTest:
White
F-statistic
6.658663
Prob.F(2.34)
0.0036
Obs*R-squared
10.41354
ProbChi-Square
(2)
0.0055
ScaieCexplainedSS
10,36651
Prob.ChkSquare{2)
0,0056
TestEquation:
DependentVariaDle:
RESIDA2Method:
LeastSquares
Date:
09/07/15Time:
11:
02
Sample:
19782014
Irt匚ludedobservations:
37
CoefficientStd.Errort-StatisticProb.
C
0.039819
0.0458S60.867779
0.3916
URIR
-0.042859
0038582-1110863
0.2744
URIRA2
0.011780
0.0079411.483508
0.1472
R-squared
0.281447
Meandependentvar
0.008786
AdjustedR-squared
0239179
S.D.dependentvar
0013287
S.Eofregression
0.011590
Akaikeinfocriterion
-5999002
Sumsquaredresid
0.004567
£匚hwamcriterion
-5.869187
Loglikelihood
1139963
Hannan*Quinnenter.
-5953754
F-statistic
6.658663
Durbin-Watsonstat
0.374222
Prob(F-statistic)
0.003629
十五、WLS加权最小二乘法
DependentVariable:
UR
Method:
LeastSquaresDate:
09/07/15Time:
11:
29
Sample:
19782014includedobservations:
37
Weightingseries:
W
Coefficient
Std.Error
t-Statistic
Prob.
URIR
0.168557
0.031139
5,413033
0.0000
C
-0.086069
0.076542
-1.124465
0.2685
WeightedStatistics
R-squared
0.455684
(Jeandependentvar
0.328603
AdjustedR-squared
0440132
S.Ddependentvar
0.100794
SE,ofregression
0.090040
Akaikeinlacriterion
*1.924588
Sumsquaredresid
0.283752
Schwarzcriterion
-1,837512
LogliKelitiaod
37.50488
Hannan-Quinncriter
-1893890
F-statistic
29.30092
Durbin-Watsonstat
0.279531
Prob(F-statistic)
0.000005
UnweightedStatisti匚s
R-squared
□378738
Meandependentvar
0.33614S
AdjustedR-squared
0.350987
SD.dependentvar
0.120526
S.E.ofregression
0.096427
Sumsquaredresid
0.325433
Durbin-Watsonstat
0.310271
十六、残差自相关图及其Q检验统计量
CorrelogramofResiduals
Date:
09/07/15Time:
15:
50
Sample:
19782014
Includedobservations:
37
AutocorrelationPartialCorrelationACPACQ-StatProb
1
1
1
1
1
□1
10.7120.71220.3330000
205730133338500000
1
二
1匚
II
30.380-014239.9630.000
1
1[
1
40.239-005942.4520.000
II
]1
1(
1
50.111-0.05543.0060000
II
】>
H
]1
60.0600.05243.1760.000
1
]>
H
]1
70.0530.07843.3100.000
II
]>
H
]1
&0.0680.04243.5390.000
II
]>
1(
II
90.056-0.05143.6980.000
1
1>
1
]1
100.0340.05044.0750000
1
11
1
]1
110.11S008244.3430000
1
□1
1
11
120.157007246.2730000
1
n>
1
:
1
130.217011449.1000.000
II
ZI|
1[
1
140.203-0.06751.6340.000
1
□>
1匚
1
150148-0.11953.1140.000
1
1
1匚
1
150024-0.17153.1520.000
1-16阶的p值都小于0.01,说明拒绝原假设,残差序列存在自相关性。
十七、残差自相关LM检验结果
Breuscn-GocrrreyserialCorrslatJonLMT©st:
F-statistic
Otl£*R-£.口1_1自「耳€]
N石4640•尹
Rrot>F<2,33)
Pro.Chi-Scilj岂
DOOOO
0.0000
TestEqliation:
DepenitjentVaria&le'RSSIO
Metnodii_eastSquares
Date:
09/07/15Time:
1
Sampile:
19ZB201^
lindudiedolbservations.37
PreSainnp0&llue1日口些旦日「孕导iClinH导settozi^ro
Coefficient
Std.Error
t-Statisti匚
Proti.
URIRo
RESIDC-1)
RESIDC-2)
-O„O49356O-12B67OO.63&8SSOJ284L8O9
0.02:
3318OO日口日64Q.16S1Z3
Ck"4374
-2.0724212
2"40IS日
3.92OB26
163332J2
0.0461O04-2^0.0004O-11il9
R-SQuarecl
AdjustedR-squarecdlS.E.ofreoresslionSLimscjuoreclresidlLoo111h(□F-statistic
Prob(F-statistic>
O„61S958
OS81045
0.061509
O.121^852
S2.79260
64271
O.DDOOO1
MeandependentvarS.DdependentvarAKalKeinTocriterionSchwar-z.criterion
M^nin^n-Ouinn:
crirtierDurbin-Watson
O0-95029-2.637-453-2„4I333OO-2.S76OS6
1„2O4229
F与Obs两个的P值显示:
存在自相关
十八、Newey-West—致协方差估计
Dependentvariable:
UR
Method:
LeastSquares
Date:
09/07/15Time:
1635
Sample:
19782014
Includedobservations:
37
NeweyAVestHACStandardErrors&Covariance(lagtruncation=3)
Coefficient
StdLErrort-Statistic
Prob.
URIR
0.161922
0.0508133,186620
0.0030
C
-0.069449
0.109907-0.631837
0.5316
R-squared
0.379375
Meandependentvar
0.336145
AdjustedR-squared
0.361643
S.Ddependentvar
0.120626
S.E.ofregression
0.096377
Akaikeinfocriterion
-1.788557
Sumsquaredresid
0.325099
Schwarzcriterion
-1,701481
Loglikelihood
35.08831
Hannan-Quinncriter.
*1.757959
F-statistic
2139474
Durbin-Watsonstat
0.289676
Prob(F-statistic)
0.000049
十九、两阶段TSLS估计检验结果
DependentVariable:
URl^ethod:
Two-StageLeastSquares
Date:
09/07/15Time:
16:
43
Sample:
19782014
Indudedobservations:
37
InstrumentlistCUR
Coefficient
Std.Errort-Statistic
Prob.
URIR
0.426813
0.0922754.625445
0.0000
C
*0732965
0.232564-3.151675
0.0033
R-squared
-0.635916
Meandependentvar
0.336145
AdjustedR-squared
-0.682657
S.D.dependentvar
0.120626
S.E,ofregression
0.156473
Sumsquaredredid
0856934
F-statistic
2139474
Durbin-Watsonstat
0.632913
Prob(F-statistic)
0.000049
SecontTStageSSR
4.14E*30
二十、广义矩估计GMM检验结果
toependentVariable:
UR
Method:
GeneralizedMethodofMoments
Date:
09/07/15Time:
16:
51
Sample:
19782014
Includedobservations:
37